AQUACULTURE ENVIRONMENT INTERACTIONSAquacult Environ Interact

Vol. 12: 429–445, 2020https://doi.org/10.3354/aei00376

Published October 22

1. INTRODUCTION

Atlantic salmon Salmo salar aquaculture is of inter-national socioeconomic importance, and the process

of domestication has resulted in significant phenotypic(i.e. physiological, Handeland et al. 2003; be hav ioural,Fleming et al. 1996; morphological, Fleming et al.1994); and genetic (Cross & King 1983, Karlsson et al.

© The authors 2020. Open Access under Creative Commons byAttribution Licence. Use, distribution and reproduction are un -restricted. Authors and original publication must be credited.

Publisher: Inter-Research · www.int-res.com

*Corresponding author: ibradbur@me.com

REVIEW

Beyond hybridization: the genetic impacts ofnon-reproductive ecological interactions of

salmon aquaculture on wild populations

I. R. Bradbury1,*, I. Burgetz2, M. W. Coulson2, E. Verspoor3, J. Gilbey4, S. J. Lehnert1, T. Kess1, T. F. Cross5, A. Vasemägi6,7, M. F. Solberg8, I. A. Fleming9, P. McGinnity5

1Fisheries and Oceans Canada, Northwest Atlantic Fisheries Centre, 80 E White Hills Rd, St. John’s, Newfoundland, A1C 5X1, Canada

2Fisheries and Oceans Canada, 200 Kent Street Ottawa, Ontario, K1A 0E6, Canada3Rivers and Lochs Institute, University of the Highlands and Islands, Inverness, IV3 5SQ, UK

4Marine Scotland Science, Freshwater Fisheries Laboratory, Pitlochry, Scotland PH16 5LB, UK5School of Biological, Earth and Environmental Sciences, University College, Cork, T12 YN60, Ireland

6Chair of Aquaculture, Institute of Veterinary Medicine and Animal Sciences, Estonian University of Life Sciences,51014 Tartu, Estonia

7Department of Aquatic Resources, Institute of Freshwater Research, Swedish University of Agricultural Sciences,17893 Drottningholm, Sweden

8Population Genetics Research Group, Institute of Marine Research, 5817 Bergen, Norway9Memorial University of Newfoundland, Department of Ocean Sciences, St. John’s, Newfoundland, A1C 5S7, Canada

ABSTRACT: Cultured Atlantic salmon Salmo salar are of international socioeconomic value, andthe process of domestication has resulted in significant behavioural, morphological, and allelic dif-ferences from wild populations. Substantial evidence indicates that direct genetic interactions orinterbreeding between wild and escaped farmed Atlantic salmon occurs, genetically altering wildsalmon and reducing population viability. However, genetic interactions may also occur throughecological mechanisms (e.g. disease, parasites, predation, competition), both in conjunction withand in the absence of interbreeding. Here we examine existing evidence for ecological and non-reproductive genetic interactions between domestic Atlantic salmon and wild populations and thepotential use of genetic and genomic tools to resolve these impacts. Our review identified examplesof genetic changes resulting from ecological processes, predominately through pathogen or para-site transmission. In addition, many examples were identified where aquaculture activities haveeither altered the selective landscape experienced by wild populations or resulted in reductions inpopulation abundance, both of which are consistent with the widespread occurrence of indirectgenetic changes. We further identify opportunities for genetic or genomic methods to quantify theseimpacts, though careful experimental design and pre-impact comparisons are often needed toaccurately attribute genetic change to aquaculture activities. Our review indicates that ecologicaland non-reproductive genetic interactions are important, and further study is urgently needed tosupport an integrated understanding of aquaculture–ecosystem interactions, their implications forecosystem stability, and the development of potential mitigation and management strategies.

KEY WORDS: Atlantic salmon · Aquaculture · Management · Genetic

OPENPEN ACCESSCCESS

2011, Wringe et al. 2019) differences from wild popu-lations. Escape events from Atlantic salmon net penaquaculture are a regular occurrence (Keyser et al.2018), and the number of escapees can equate to anappreciable fraction of, or exceed, wild Atlantic salmoncensus size (Morris et al. 2008, Skilbrei et al. 2015,Wringe et al. 2018). There is substantial evidence thatdirect genetic interactions, defined as interbreeding,occurs between wild Atlantic salmon and escaped do-mestic individuals (Karlsson et al. 2016, Glover et al.2017, Wringe et al. 2018) and can genetically alterwild salmon and reduce population viability (McGin-nity et al. 2003, Bourret et al. 2011, Glover et al. 2013,Bolstad et al. 2017, Bradbury et al. 2020). Both in Can-ada and Norway, recent evidence suggests hybridiza-tion may be extensive following escape events (Karls-son et al. 2016, Wringe et al. 2018) and accounts for asubstantial proportion of production in smaller rivers(Syl vester et al. 2018b). Accordingly, escaped farmedsal mon and direct genetic interactions have beenidentified as a major threat to the per-sistence and stability of wild Atlanticsalmon across the North Atlan tic (For -seth et al. 2017, Bradbury et al. 2020).

However, genetic impacts may alsooc cur, either in concert with or in theab sence of hybridization (Verspoor etal. 2015), due to ecological interactionssuch as competition, predation, anddisease or parasite transfer. These non-reproductive genetic changes in wildpopulations can result from ecologicalchanges that either alter the selectivelandscape experienced by native fish,and thus change allele frequencies ofloci linked to fitness, and/ or reduce pop-ulation abundance, re sulting in a lossof genetic diversity (Fig. 1). As these ef-fects do not involve hybridization, theycan arise whether domestic animals es-cape or remain in containment and im-pact wild populations of any nativespecies. Although practices to limit re-productive genetic inter actions withwild Atlantic salmon have been imple-mented in many areas through the useof sterilization (Verspoor et al. 2015), ex-otic species, and improved containmentstrategies (Dise rud et al. 2019), theseefforts do not prevent non-reproductivegenetic effects. In other species such asbrown trout Salmo trutta or Pacificsalmon species (Onco rhynchus spp.)

where hybrid ization with escapees is not common orpossible, ecologically induced genetic interactionswith Atlantic salmon aquaculture re main an ongoingconcern (e.g. Coughlan et al. 2006, Ford & Myers2008). Moreover, given recent trends in industry ex-pansion (e.g. DFO 2016, 2018) and growing concernsregarding the amplification of pests and pathogenssuch as sea lice through net pen aquaculture (e.g.Vollset et al. 2016, Karbowski et al. 2019), the potentialfor both ecological and non-reproductive genetic in-teractions is likely to in crease. Nonetheless, despitethe potentially broad reaching and significant impactsof non-reproductive genetic interactions on wild At-lantic salmon and other species, the evidence for theirpresence and our ability to quantify their magnitudehas been limited to date (Verspoor et al. 2015).

The goal of this review is to highlight evidence per-taining to the potential for ecological and associatednon-reproductive genetic impacts of Atlantic salmonaquaculture on wild populations. Specifically, our

430 Aquacult Environ Interact 12: 429–445, 2020

Fig. 1. Schematic of reproductive and non-reproductive genetic interactions be tween wild and domestic Atlantic salmon Salmo salar

ob jectives are to (1) review examples of geneticchanges in wild populations resulting from eco -logical interactions, or likely more common, evi-dence for changes in population abundance or theenvironment experienced by wild populations; and(2) discuss the opportunity recent advances in popu-lation genomic approaches present for the assess-ment of these genetic impacts. Through our review,we highlight opportunities for the further study ofnon- reproductive genetic impacts of Atlantic salmonaquaculture on wild populations. We directly buildon previous reviews and empirical studies focusingon hybridization and introgression (e.g. Karlsson etal. 2016, Glover et al. 2017, Bradbury et al. 2020) andon risk assessments considering both reproductiveand non-reproductive effects (e.g. Verspoor et al.2015). Ultimately, we suggest that ecological andsubsequent non-reproductive genetic impacts arelikely ubiquitous wherever salmon farming occurs,and that further research is urgently required to bet-ter understand the magnitude of these interactionsand provide advice regarding impact managementand mitigation.

2. EVIDENCE FOR ECOLOGICAL ANDNON-REPRODUCTIVE GENETIC IMPACTS

Atlantic salmon net pen aquaculture represents asubstantial change to the natural environment andthus the adaptive landscape experienced by wild in-dividuals (Garcia de Leaniz et al. 2007). As such, itcan alter the stability and future evolutionary trajec-tories of wild populations. Furthermore, it might beex pected that adjustments to a new adaptive land-scape will result in reductions in productivity throughincreased maladaptation predicted by theoretical demographic-evolutionary models (Bürger & Lynch1995, Gomulkiewicz & Holt 1995, Kirkpatrick & Bar-ton 1997). Existing studies address genetic changesin naïve populations through disease and parasitetransmission, the potential for recovery of disease orparasite resistance through natural selection, obser-vations on genetic changes in co-occurring congenerspecies, and impacts of the farming of non-nativespecies or subspecies. Examples of the latter are thefarming of European origin salmon on both the eastand west coasts of North America as well as in west-ern South America or Australia. Below we review theliterature related to non-reproductive genetic inter-actions associated with disease and parasite transfer,increased predation pressure, and finally, increasedcompetition (see Table 1). In each case, we first high-

light examples of genetic change re sulting from theseinteractions and then set out evidence of demo -graphic decline or the potential for selection consis-tent with significant genetic impacts. In practice, itcan be difficult to distinguish the im pacts of repro-ductive and non-reproductive genetic interactions inexamples related to wild Atlantic salmon. As such,here we focus on instances where mechanisms havebeen identified which are clearly non-reproductive innature.

2.1. Ecological and non-reproductive geneticchanges through disease transmission

Ecological and genetic interactions via diseasetransmission may result in both alterations to the se -lective landscape potentially impacting immune as -sociated genetic variation as well as reductions inoverall genetic diversity due to demographic decline.To date, few studies have examined the presence ofgenetic changes due to disease transfer (Table 1A).However, de Eyto et al. (2007, 2011) present evidenceof genetic impacts due to novel disease exposure as-sociated with aquaculture activities. In these studies,the progeny of Atlantic salmon from a river withoutprevious exposure to aquaculture were trans ferred toa river with a long history of associated farming andcaptive breeding that was expected to have acquirednovel micro- and macro-parasitic communities. Thisexperimental design enabled the exposure of animalsto novel disease challenges associated with escapesor inadvertent or deliberate introductions. Compari-son of observed and expected genotype frequenciesat a marker locus for the MHC class II alpha gene andcontrol neutral microsatellite loci of parr and migrantAtlantic salmon stages in the wild demonstrated thatgenetic change had occurred, and that selection waslikely a result of disease-mediated natural selection,rather than any demographic event.

A substantial and growing body of research sup-ports the hypothesis that wild salmon populations areadapted to local pathogen communities both in spaceand time (Dionne et al. 2007, Tonteri et al. 2010, Con-suegra et al. 2011, Kjærner-Semb et al. 2016, Prit -chard et al. 2018, Zueva et al. 2018). This suggests agenetic basis for differences in population immunityand that the introduction of new pathogens into sus-ceptible populations could both impose novel selec-tion pressures and reduce genetic diversity throughdemographic decline. The possibility that pathogentransfer from domestic to wild salmon could drivegenetic change in wild populations is supported by

431Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

432 Aquacult Environ Interact 12: 429–445, 2020

Inte

ract

ion

Pri

mar

y ob

serv

atio

nE

vid

ence

(d

irec

tor

su

pp

orti

ve)

Sel

ecti

on/

dem

ogra

ph

icS

pec

ies

imp

acte

dR

efer

ence

(A)

Dis

ease

tra

nsf

erC

omm

on g

ard

en e

xper

i-m

ent

(naï

ve n

on-l

ocal

wild

pop

ula

tion

intr

odu

ced

into

dif

fere

nt

rive

r sy

stem

as

egg

s)

Evi

den

ce o

f al

lele

fre

qu

ency

ch

ang

e at

maj

or h

isto

com

pat

ibili

ty (

MH

) g

enes

du

rin

g f

irst

6 m

onth

s in

intr

odu

ced

pop

ula

tion

; no

chan

ge

in lo

cal p

opu

lati

on)

Su

pp

orti

veS

elec

tion

Atl

anti

c sa

lmon

S

. sal

ard

e E

yto

et a

l. (2

007)

Com

mon

gar

den

exp

eri-

men

t (n

aïve

non

-loc

al w

ildp

opu

lati

on in

trod

uce

d in

tod

iffe

ren

t ri

ver

as e

gg

s)

Evi

den

ce o

f d

iffe

ren

t al

lele

fre

qu

ency

chan

ge

at m

ajor

his

toco

mp

atib

ility

(M

H)

gen

es in

intr

odu

ced

pop

ula

tion

fro

m 6

mon

ths

to 1

8 m

onth

s; n

o ch

ang

e in

loca

lp

opu

lati

on)

Su

pp

orti

veS

elec

tion

Atl

anti

c sa

lmon

S.s

alar

de

Eyt

o et

al.

(201

1)

Gen

etic

su

rvey

of

nat

ura

lp

opu

lati

ons

(not

ass

ocia

ted

wit

h a

qu

acu

ltu

re)

Evi

den

ce o

f cl

inal

geo

gra

ph

ical

res

pon

sein

maj

or h

isto

com

pat

ibili

ty (

MH

) g

enes

inre

spon

se t

o w

ater

tem

per

atu

re v

aria

tion

)

Su

pp

orti

veS

elec

tion

Atl

anti

c sa

lmon

S.s

alar

Dio

nn

e et

al.

(200

7)

Gen

etic

su

rvey

of

nat

ura

lp

opu

lati

ons

pot

enti

ally

sen

siti

ve a

nd

tol

eran

t of

Gyr

odac

tylu

s sa

lari

s

Evi

den

ce o

f cl

inal

geo

gra

ph

ical

res

pon

sein

maj

or h

isto

com

pat

ibili

ty (

MH

) &

oth

erim

mu

ne

rele

van

t g

enes

in r

esp

onse

to

wat

er t

emp

erat

ure

var

iati

on)

Su

pp

orti

ve(p

ossi

ble

dir

ect

link

to

G. s

alar

isp

aras

ite)

Sel

ecti

onA

tlan

tic

salm

onS

.sal

arT

onte

ri e

t al

. (20

10)

Gen

etic

su

rvey

of

nat

ura

lp

opu

lati

ons

in a

reas

wit

han

d w

ith

out

aqu

acu

ltu

reac

tivi

ty

Evi

den

ce o

f sp

atia

l alle

le v

aria

tion

at

maj

or h

isto

com

pat

ibili

ty (

MH

) g

enes

Su

pp

orti

ve(p

ossi

ble

dir

ect

link

to

vira

lp

ath

ogen

s)

Sel

ecti

onA

tlan

tic

salm

onS

.sal

arC

onsu

egra

et

al. (

2011

)

Gen

etic

su

rvey

of

nat

ura

lp

opu

lati

ons

in r

egio

n o

fsi

gn

ific

ant

aqu

acu

ltu

reac

tivi

ty

Evi

den

ce o

f S

NP

var

iati

on a

ssoc

iate

d w

ith

sele

ctiv

e sw

eep

s of

imm

un

e re

spon

seg

enes

Su

pp

orti

ve(s

ourc

e of

sele

ctiv

e ag

ent

un

kn

own

)

Sel

ecti

onA

tlan

tic

salm

onS

.sal

arK

jærn

er-S

emb

et

al. (

2016

)

Gen

etic

su

rvey

of

nat

ura

lp

opu

lati

ons

wit

hin

sin

gle

larg

e ri

ver

com

ple

x (n

otas

soci

ated

wit

h a

qu

acu

ltu

re)

Evi

den

ce o

f S

NP

var

iati

on a

ssoc

iate

d w

ith

maj

or h

isto

com

pat

ibili

ty (

MH

) g

enes

Su

pp

orti

veS

elec

tion

Atl

anti

c sa

lmon

S.s

alar

Pri

tch

ard

et

al. (

2018

)

Dis

ease

scr

een

ing

of

esca

ped

far

med

Atl

anti

csa

lmon

in a

wild

riv

er

Vir

us

infe

cted

esc

aped

far

med

sal

mon

ente

rin

g r

iver

s n

ear

cag

e si

tes

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S.s

alar

Mad

hu

n e

t al

. (20

15)

Dis

ease

scr

een

ing

of

retu

rnin

g w

ild A

tlan

tic

salm

on in

Nor

way

at

6 si

tes

Evi

den

ce f

or t

he

infe

ctio

n o

f w

ild s

alm

onfr

om e

scap

ed f

arm

ed s

alm

on a

t m

arin

efe

edin

g a

reas

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S.s

alar

Mad

hu

n e

t al

. (20

18)

Gen

etic

scr

een

ing

of

PR

V in

wild

an

d f

arm

ed A

tlan

tic

salm

on

Evi

den

ce f

or lo

ng

dis

tan

ce t

ran

smis

sion

of

PR

V li

kel

y as

soci

ated

wit

h a

qu

acu

ltu

rein

du

stry

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S.s

alar

Gar

seth

et

al. (

2013

)

Tab

le 1

. Su

mm

ary

of s

tud

ies

pre

sen

tin

g e

vid

ence

for

or

con

sist

ent

wit

h t

he

pot

enti

al f

or e

colo

gic

al a

nd

non

-rep

rod

uct

ive

gen

etic

inte

ract

ion

s am

ong

Atl

anti

c sa

lmon

S

alm

o sa

lar

aqu

acu

ltu

re a

nd

wild

sal

mon

id p

opu

lati

ons.

N/A

: not

ap

plic

able

Tab

le c

onti

nu

ed o

n n

ext

pag

e

433Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

Inte

ract

ion

Pri

mar

y ob

serv

atio

nE

vid

ence

(d

irec

tor

su

pp

orti

ve)

Sel

ecti

on /

dem

gra

ph

icS

pec

ies

imp

acte

dR

efer

ence

Rev

iew

of

stu

die

s d

ocu

-m

enti

ng

fu

run

culo

sis

pre

vale

nce

in N

orw

ay f

rom

1964

–199

2

Evi

den

ce f

or t

he

tran

sfer

of

furu

ncu

losi

sfr

om f

ish

far

ms

to w

ild s

alm

onid

s in

Nor

way

Su

pp

orti

veD

emo-

gra

ph

icV

ario

us

salm

onid

sJo

hn

sen

& J

ense

n (

1994

)

Gen

etic

scr

een

ing

of

ISA

Vva

rian

ts in

wild

an

d f

arm

edsa

lmon

in N

orw

ay

Evi

den

ce f

or t

he

hor

izon

tal t

ran

smis

sion

of I

SA

V v

aria

nts

see

n in

far

med

sal

mon

to

wild

pop

ula

tion

s

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S.s

alar

Nyl

un

d e

t al

. (20

19)

(B)

Par

asit

e tr

ansf

erS

tati

stic

al m

odel

ing

of

the

effe

ct o

n r

etu

rn r

ates

of

sea

lice

leve

ls (

low

/med

/hig

h)

over

a 2

6 ye

ar p

erio

d f

or1S

W E

rrif

f sa

lmon

Wild

sal

mon

ret

urn

s w

ere

stro

ng

lyre

du

ced

(>

50%

) fo

llow

ing

yea

rs w

ith

hig

h li

ce le

vels

du

rin

g s

mol

t ou

t-m

igra

-ti

on (

farm

s lo

cate

d a

t th

e m

outh

of

the

estu

ary)

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S. s

alar

Sh

eph

ard

& G

arg

an (

2017

)

Tag

/rec

aptu

re e

xper

imen

tof

pro

ph

ylac

tica

lly t

reat

edsm

olts

exp

osed

to

dif

fere

nt

farm

-ori

gin

sea

lice

pre

ssu

re

Rec

aptu

re r

ate

of u

ntr

eate

d a

du

lt s

alm

onfo

llow

ing

exp

osu

re t

o h

igh

sea

lice

den

sity

was

0.0

3%

com

par

ed t

o tr

eate

dsa

lmon

(1.

86%

)

Dir

ect

Bot

hA

tlan

tic

salm

onS

. sal

arB

øhn

et

al. (

2020

)

Ass

ocia

tion

bet

wee

n s

ealic

e co

un

ts o

n f

arm

edA

tlan

tic

salm

on a

nd

wild

out-

mig

rati

ng

ch

um

sal

mon

Sig

nif

ican

t p

osit

ive

asso

ciat

ion

bet

wee

nth

e se

a lic

e ab

un

dan

ce o

n f

arm

s an

d t

he

likel

ihoo

d t

hat

juve

nile

ch

um

sal

mon

wou

ld b

e in

fest

ed. I

ncr

ease

d a

bu

nd

ance

of li

ce o

n f

arm

s w

as n

ot s

ign

ific

antl

yas

soci

ated

wit

h t

he

leve

ls o

f in

fest

atio

nob

serv

ed o

n ju

ven

ile c

hu

m s

alm

on

Su

pp

orti

veB

oth

Ch

um

sal

mon

On

corh

ynch

us

ket

a

Nek

ouei

et

al. (

2018

)

Exp

erim

enta

l sea

lice

infe

ctio

n o

f w

ild b

row

n t

rou

tp

ost-

smol

ts a

nd

exa

min

a-ti

ons

of m

arin

e m

igra

tory

beh

avio

r

Exp

erim

enta

l sea

lice

infe

ctio

n a

ssoc

iate

dw

ith

incr

ease

d m

orta

lity,

an

d d

ecre

ased

mig

rati

on d

ista

nce

, an

d m

arin

e re

sid

ency

Su

pp

orti

veB

oth

Sea

tro

ut

S.t

rutt

aS

erra

-Llin

ares

et

al. (

2020

)

Rev

iew

pap

er: i

nte

gra

tin

gla

bor

ator

y an

d f

ield

ob

ser-

vati

onal

stu

die

s of

lice

on

out-

mig

rati

ng

S. s

alar

and

S.t

rutt

a

Sea

lice

load

s on

ou

t-m

igra

tin

g s

ea t

rou

tin

are

as w

ith

aq

uac

ult

ure

com

mon

lyex

ceed

th

resh

old

leve

ls t

hat

are

kn

own

to

ind

uce

ph

ysio

log

ical

com

pro

mis

e or

mor

talit

y in

lab

orat

ory

exp

erim

ents

Su

pp

orti

veB

oth

Sea

tro

ut

S.t

rutt

aT

hor

stad

& F

inst

ad (

2018

)

Rev

iew

pap

er: i

nte

gra

tin

gla

bor

ator

y an

d f

ield

ob

ser-

vati

onal

stu

die

s of

lice

on

out-

mig

rati

ng

S. s

alar

and

S.t

rutt

a

Pre

mat

ure

mig

rato

ry r

etu

rnD

irec

tD

emo-

gra

ph

icS

ea t

rou

tS

.tru

tta

Th

orst

ad &

Fin

stad

(20

18)

Tab

le 1

(con

tin

ued

)

Tab

le c

onti

nu

ed o

n n

ext

pag

e

434 Aquacult Environ Interact 12: 429–445, 2020

Inte

ract

ion

Pri

mar

y ob

serv

atio

nE

vid

ence

(d

irec

tor

su

pp

orti

ve)

Sel

ecti

on /

dem

ogra

ph

icS

pec

ies

imp

acte

dR

efer

ence

Rev

iew

pap

er: i

nte

gra

tin

gla

bor

ator

y an

d f

ield

ob

ser-

vati

onal

stu

die

s of

lice

on

out-

mig

rati

ng

S. s

alar

and

S.

tru

tta

Su

mm

ary

of m

eta-

anal

ysis

an

d t

agg

edtr

eate

d s

mol

t su

rviv

al t

o re

turn

ing

ad

ult

sex

per

imen

t

Su

pp

orti

veB

oth

Atl

anti

c sa

lmon

S. s

alar

Th

orst

ad &

Fin

stad

(20

18)

Sea

lice

ab

un

dan

ce o

n o

ut-

mig

rati

ng

pin

k s

alm

on a

nd

chu

m s

alm

on d

iffe

ren

ces

pre

- an

d p

ost-

exp

osu

re t

oA

tlan

tic

salm

on f

arm

s

Qu

anti

tati

ve e

stim

ate

of t

ran

smis

sion

rate

s fr

om f

arm

to

out-

mig

rati

ng

pin

k a

nd

chu

m s

alm

on, i

ncl

ud

ing

su

bse

qu

ent

tran

smis

sion

dyn

amic

s of

lice

wit

hin

th

ew

ild p

opu

lati

on

Su

pp

orti

veD

emo-

gra

ph

icP

ink

sal

mon

O.g

orb

usc

ha

and

ch

um

sal

mon

O.k

eta

Krk

ošek

et

al. (

2005

)

Hie

rarc

hic

al m

odel

of

stoc

k-

recr

uit

dyn

amic

s of

coh

osa

lmon

wit

h d

iffe

ren

tial

sea

lice

infe

stat

ion

Coh

o sa

lmon

pop

ula

tion

pro

du

ctiv

ity

inan

are

a of

inte

nsi

ve s

alm

on a

qu

acu

ltu

rew

as d

epre

ssed

ap

pro

xim

atel

y se

ven

fold

du

rin

g a

per

iod

of

salm

on lo

use

infe

sta-

tion

s co

mp

ared

to

un

exp

osed

pop

ula

tion

s.

Su

pp

orti

veD

emo-

gra

ph

icC

oho

salm

onO

.kis

utc

hC

onn

ors

et a

l. (2

010)

Mod

elin

g e

ffec

t of

sea

lice

infe

ctio

ns

on p

opu

lati

onab

un

dan

ce o

f p

ink

sal

mon

Pin

k s

alm

on p

opu

lati

ons

exp

osed

to

salm

on f

arm

s; m

orta

lity

rate

cau

sed

by

sea

lice

was

est

imat

ed t

o ra

ng

e fr

om 1

6 to

97%

Su

pp

orti

veD

emo-

gra

ph

icP

ink

sal

mon

O.g

orb

usc

ha

Krk

ošek

et

al. (

2007

)

An

alys

is o

f sp

awn

er-r

ecru

itd

ata

and

sea

lice

ab

un

dan

ceon

far

ms

Sea

lice

cou

nts

on

fis

h f

arm

s w

ere

neg

ativ

ely

asso

ciat

ed w

ith

ad

ult

ret

urn

sof

2 s

pec

ies

of P

acif

ic s

alm

on

Su

pp

orti

veD

emo-

gra

ph

icP

ink

sal

mon

O.g

orb

usc

ha

and

Coh

o sa

lmon

O. k

isu

tch

Krk

ošek

et a

l. (2

011a

)

Scr

een

ing

of

pyr

eth

roid

resi

stan

ce g

enot

ype

inL

epeo

ph

thei

rus

salm

onis

over

tim

e

Wid

esp

read

ch

ang

es in

th

e fr

equ

ency

of

gen

otyp

e as

soci

ated

wit

h p

yret

hro

idre

sist

ance

in s

ea li

ce a

cros

s th

e N

orth

Atl

anti

c

Dir

ect

Sel

ecti

onS

alm

on lo

use

Lep

eop

hth

eiru

ssa

lmon

is

Bør

retz

en F

jørt

oft

et a

l.(2

020)

G. s

alar

isin

fect

ion

ass

oci-

ated

wit

h w

ild s

alm

onp

opu

lati

on d

eclin

e

Wild

sto

cks

dec

reas

ed in

siz

e b

y an

aver

age

of 8

5%

an

d s

mol

t n

um

ber

sd

ecre

ased

by

as m

uch

as

98%

fol

low

ing

intr

odu

ctio

n o

f G

. sal

aris

into

Nor

way

Su

pp

orti

veD

emo-

gra

ph

icA

tlan

tic

salm

onS

. sal

arD

enh

olm

et

al. (

2016

)

Gen

omic

bas

is o

f re

sist

ance

to G

. sal

aris

Iden

tifi

ed 3

gen

omic

reg

ion

s as

soci

ated

wit

h a

dap

tati

on t

o p

aras

ite

resi

stan

ce in

wild

sal

mon

Su

pp

orti

veN

/AA

tlan

tic

salm

onS

. sal

arZ

uev

a et

al.

(201

4)

Gen

omic

bas

is o

f re

sist

ance

to G

. sal

aris

Iden

tifi

ed 5

7 ca

nd

idat

e g

enes

pot

enti

ally

un

der

pos

itiv

e se

lect

ion

ass

ocia

ted

wit

hG

. sal

aris

resi

stan

ce a

nd

en

rich

ed f

orly

mp

h n

ode

dev

elop

men

t, f

ocal

ad

hes

ion

gen

es a

nd

an

ti-v

iral

res

pon

ses

Su

pp

orti

veN

/AA

tlan

tic

salm

onS

. sal

arZ

uev

a et

al.

(201

8)

Tab

le 1

(con

tin

ued

)

Tab

le c

onti

nu

ed o

n n

ext

pag

e

several recent findings documentingthe potential for exposure and support-ing pathogen transfer as mechanismsfor genetic impacts (Table 1A). First,Madhun et al. (2015) report the detec-tion of virus infected escaped farmedsalmon entering rivers near cage sites,suggesting clear evidence of exposureof freshwater rearing juvenile salmonpopulations to aquaculture associatedpathogens. Second, Madhun et al. (2018)also document the presence of piscineortho reo virus (PRV) in returning wildadult Atlantic salmon in Norway, andthat the frequency of infection in -creased with body size and displayedno geographic signal, suggesting infec-tion was occurring between escapeesand wild salmon at marine feedingareas. Nylund et al. (2019) report thatinfectious sal mon anemia virus (ISAV)variants in farmed sal mon are increas-ing in prevalence in the wild consis-tent with horizontal transmission fromfarmed sal mon to wild populations.Similarly, Garseth et al. (2013) examinepathogen transfer between wild andfarmed salmon using analysis of proteincoding sequences in PRV in Norwayand suggest occurrence in the wild isdue to long distance transmission likelyassociated with the aquaculture indus-try. Finally, several studies have docu-mented the spread of furun cu losis, asepticemic bacterial disease, from fishfarms to wild salmonids in Norwegianrivers (John sen & Jensen 1994). Takentogether, these findings indicate thatecologically induced genetic im pactson wild salmon populations associatedwith disease transmission from aqua-culture populations are highly likely.However, both the magnitude of newselection pressures and demographicimpacts are uncertain and likely casespecific.

Diseases, introduced or increased inincidence by salmon aquaculture activ-ities, could also have an impact on co-occurring wild species such as anadro-mous brown trout, as implied by thesteep decline in anadromous trout num-bers in many Irish, Scottish, and Nor-

435Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

Inte

ract

ion

Pri

mar

y ob

serv

atio

nE

vid

ence

(d

irec

tor

su

pp

orti

ve)

Sel

ecti

on /

dem

ogra

ph

icS

pec

ies

imp

acte

dR

efer

ence

Gro

wth

an

d s

urv

ival

of

sea

lice

infe

cted

Arc

tic

Ch

arr

Infe

ctio

n in

ten

sity

cor

rela

ted

pos

itiv

ely

wit

h m

orta

lity

and

neg

ativ

ely

wit

h g

row

thin

exp

erim

enta

l tri

als

Su

pp

orti

veB

oth

Arc

tic

char

r S

alve

lin

us

alp

inu

s

Fje

lldal

et

al. (

2019

)

(C)

Pre

dat

ion

Incr

ease

d p

red

atio

n o

n w

ildsp

ecie

sIn

crea

sed

avi

an p

red

atio

n o

n w

ild s

alm

onan

d b

row

n t

rou

t fo

llow

ing

th

e re

leas

e of

cap

tive

bre

d s

mol

ts

Su

pp

orti

veD

emo-

gra

ph

ic /

sele

ctiv

e?

Bro

wn

tro

ut

S.t

rutt

aK

enn

edy

& G

reer

(19

88)

Pre

dat

ion

on

rel

ease

dfa

rmed

esc

apes

Hig

h le

vels

of

pre

dat

ion

on

rel

ease

dfa

rmed

Atl

anti

c sa

lmon

nea

r ca

ge

site

sS

up

por

tive

Dem

o-g

rap

hic

/se

lect

ive?

Atl

anti

c sa

lmon

S. s

alar

Ham

oute

ne

et a

l. (2

018)

(D)

Co

mp

etit

ion

Com

pet

itio

n b

etw

een

wild

and

far

med

juve

nile

Atl

anti

csa

lmon

in f

resh

wat

er

30%

red

uct

ion

in w

ild p

opu

lati

onp

rod

uct

ivit

y in

th

e p

rese

nce

of

farm

edfi

sh

Su

pp

orti

veD

emo-

gra

ph

icA

tlan

tic

salm

onS

. sal

arF

lem

ing

et

al. (

2000

)

Com

pet

itio

n b

etw

een

wild

and

far

med

juve

nile

Atl

anti

csa

lmon

in f

resh

wat

er

Ove

rlap

in d

iet

amon

g t

ypes

of

cros

ses

dem

onst

rate

s co

mp

etit

ion

Su

pp

orti

veD

emo-

gra

ph

icA

tlan

tic

salm

onS

. sal

arS

kaa

la e

t al

. (20

12)

Met

abol

ic r

ate

and

su

rviv

alof

far

med

Atl

anti

c sa

lmon

offs

pri

ng

Pre

sen

ce o

f w

ild-f

arm

ed h

ybri

ds

red

uce

dsu

rviv

al o

f w

ild in

div

idu

als

Su

pp

orti

veD

emo-

gra

ph

icA

tlan

tic

salm

onS

. sal

arR

ober

tsen

et

al. (

2019

)

Tab

le 1

(con

tin

ued

)

wegian rivers since the late 1980s, which may belinked to sea lice infestations (see Section 2.2) associ-ated with marine salmonid farming. A study byCough lan et al. (2006) in some Irish rivers suggestedthat salmon farming and ocean ranching could indi-rectly affect, most likely mediated by disease, the ge-netics of cohabiting anadromous brown trout by re-ducing variability at major histocompatibility class Igenes. A significant decline in allelic richness andgene diversity at the Satr-UBA marker locus, observedsince aquaculture started, which may indicate a se-lective response, was not reflected by similar reduc-tions at neutral loci. Subsequent recovery of variabilityat the Satr-UBA marker, seen among later samples,may reflect an increased contribution by residentbrown trout to the remaining anadromous population.Similarly, Miller et al. (2011) link genomic profilesconsistent with viral infection with increased likeli-hood of mortality prior to spawning in Fraser Riversockeye salmon Oncorhynchus nerka. Morton et al.(2017) document piscine ortho reo virus (PRV) in 95%of farmed Atlantic salmon in British Columbia, Can-ada, and infection rates in wild Pacific salmon of 37−45% near salmon farms, and of 5% at sites distant tofarms suggesting PRV transfer is occurring from sal -mon farms to wild salmon populations.

2.2. Ecological and non-reproductive geneticeffects through parasites

Like disease transfer, the introduction of novel par-asites could both impose new selection pressures anddrive demographic decline. Although no examples ofgenetic change attributable to parasite transfer fromsalmon aquaculture were identified, substantial re -search has demonstrated the (1) transfer of parasitesfrom aquaculture salmon to wild populations, (2) sig-nificant demographic impacts resulting, and (3) agenetic basis to resistance, all of which support thepresence of genetic change occurring as a result.Examples to date have most notably been via infec-tions of sea lice or the monogenetic trematode Gyro -dac tylus salaris (Table 1B). Declines in wild stocksattributed to sea lice outbreaks in farm-intensiveareas have been documented in Ireland, Scotlandand Norway. Thorstad & Finstad (2018) reviewed theliterature related to sea lice impacts on wild stocksdocumenting 12−29% fewer returning adult spawn-ers due to lice-induced mortality from fish farms. Inone of the most extreme cases documented to date,Shephard & Gargan (2017) suggested that one-sea-winter (1SW) salmon returns on the River Erriff were

more than 50% lower in years following high licelevels on nearby farms. This increased mortality wasin addition to decreased returns due to poorer marinesurvival. Similarly, Bøhn et al. (2020) tagged and re -leased Atlantic salmon smolts both with a prophylac-tic treatment against lice and without such treat-ment, and recaptured survivors returning to fresh-water after spending 1−4 yr at sea. They report thatthe mortality of untreated smolts was as much as 50times higher compared to treated smolts during sealice outbreaks. It is worth noting that these estimatesof lice-induced mortality among Atlantic salmonshould be considered as minimum estimates for spe-cies such as anadromous brown trout, whose marinemigrations are more coastal, thus increasing theirexposure to net pen sites (Thorstad & Finstad 2018).Recent work by Serra-Llinares et al. (2020) re portsincreased mortality, reduced marine migrations, andreduced marine residency in brown trout experimen-tally infested with sea lice, consistent with significantdemographic impacts of sea lice infection in browntrout. Similarly, for migratory Arctic char Salvelinusalpinus exposed to elevated sea lice burden due tofish farming activity (Bjørn et al. 2001), the negativeimpact on growth and survival may potentially leadto selection against anadromy (Fjelldal et al. 2019).

In addition to potential impacts on Atlantic salmo -nids, evidence also exists that the transfer of sea licefrom farmed Atlantic salmon to Pacific salmon spe-cies occurs (e.g. Nekouei et al. 2018), again consis-tent with the potential non-reproductive geneticinter actions. For example, out-migrating juvenilepink salmon O. gorbuscha and chum salmon O. keta,are estimated to experience 4 times greater sea liceinfection pressure near Atlantic salmon farms com-pared to background infection levels (Krkošek et al.2005), and in juvenile sockeye salmon O. nerka, in -fection rates were elevated after migration past thesesalmon farms (Krkošek et al. 2005, Price et al. 2011).For Coho salmon O. kisutch, ecological interactionswith infected species, as well directly with Atlanticsalmon farms, can result in higher infection levels(Connors et al. 2010). These lice infections in Pacificsalmon species have also been associated with popu-lation declines. Krkošek et al. (2007) found that sealice infestation from Atlantic salmon farms on out-migrating pink salmon smolts have led to de clines inwild populations in the Broughton Archipelago, withforecasting models suggesting that local extinctionwas imminent. For these pink salmon populations ex -posed to salmon farms, mortality rate caused by sealice was estimated to range from 16 to 97% (Krko šeket al. 2007), and population declines were also ob -

436 Aquacult Environ Interact 12: 429–445, 2020

served in Coho salmon populations (Connors et al.2010). Krkošek et al. (2011a) demonstrated that sealice abundance on fish farms in British Columbia,Canada, were negatively associated with nearby re -turns of both pink salmon and Coho salmon. Further-more, changes in parasite management on salmonfarms have been shown to help reduce infection rateson wild salmon (Peacock et al. 2013), supporting thislinkage and suggesting mitigation might be possible.

Given evidence of significant sea lice associateddemographic declines, it seems likely that sea lice-induced mortality could drive reductions in geneticdiversity. However, a large body of research suggestsresistance to sea lice may have a genetic basis and beheritable (Tsai et al. 2016, Correa et al. 2017, Robledoet al. 2019), making it highly likely that wild popula-tions would change in response to new selectionpressures. In support of this hypothesis, BørretzenFjørtoft et al. (2020) documented large-scale geneticchanges in sea lice in response to chemotherapeu-tant usage across the North Atlantic. They observedsignificant temporal changes in wild sea lice popula-tions in the frequency of a genotype associated withpyrethroid resistance due to strong selection pres-sure associated with its usage in Atlantic salmonaquaculture. Similarly, Dionne et al. (2009) reportedsignificant changes in myxozoan resistance associ-ated MHC alleles in Atlantic salmon, most likelylinked with an infection-related mortality event, fur-ther supporting the potential for parasite-associatedgenetic impacts in wild populations.

The first appearance of G. salaris in Norway hasbeen linked to the introduction of Atlantic salmonfrom Baltic catchments, resulting in high levels ofmortality among wild populations (Johnsen & Jensen1991). Admittedly, the spread of G. salaris in the wilddoes not seem primarily linked to salmon aquaculture.Instead, the transfer of individuals associated withstocking activities seems to have played a dominantrole in transmission. Nonetheless, it is included here,as it clearly illustrates the potential for the introductionof non-native individuals to transfer parasites to localpopulations, the potential for subsequent significantdemographic impacts, and a genetic basis to parasiteresistance. In G. salaris infections, very high rates ofmortality in naïve wild populations strongly supportsthe potential for significant demographic decline,losses of genetic diversity, and parasite driven selec-tion, as has been recently concluded (Karlsson et al.2020). For example, following several independent in-troductions of G. salaris into Norway, exposed wildpopulations decreased in abundance by an average of85%, and smolt numbers decreased by as much as

98% (Denholm et al. 2016). Several studies suggest agenetic basis to G. salaris resistance among wild sal -mon populations in Europe. Gilbey et al. (2006) iden-tified 10 genomic regions associated with hetero-geneity in both innate and acquired resistance usingcrosses of resistant Baltic and susceptible Atlanticpopulations. Zueva et al. (2014) compared Baltic andAtlantic At lan tic salmon populations characterized bydifferent levels of resistance to G. salaris and identi-fied 3 genomic regions potentially experiencing para-site-associated adaptation in the wild. More recently,Zueva et al. (2018) compared salmon populationsfrom northern Europe classified as extremely suscep-tible or resistant to G. salaris. They identify 57 candi-date genes potentially under resistance-associatedselection and this set of loci was shown to be enrichedfor genes associated with both innate and acquiredimmunity. These findings suggest that ecological andnon-reproductive genetic impacts on wild populationsassociated with parasite transmission, such as sea licefrom aquaculture installations, are highly likely, bothbecause of the potential for substantial mortality tooccur through exposure and for it to be selectivethrough a clear genetic basis to population differencesin resistance.

2.3. Ecological and non-reproductive geneticeffects through predation

Increased predation associated with salmon aqua-culture activities could result in both declines inabundance and selective mortality. Although directestimates are lacking, some evidence exists to sup-port the possibility of such a link, most likely throughpredators being attracted to aquaculture activities(Table 1C). Aquaculture sites have been shown to at -tract wild fish, invertebrates, marine mammals, andbirds, likely due to the addition of food, and thefarmed salmon themselves (see review in Callier etal. 2018), and the end result may be increased preda-tion on wild individuals in the vicinity. Although it ispossible that escapees could distract predators andreduce predation on wild populations through pred-ator swamping, there is no evidence to date to sup-port this. In fact, Kennedy & Greer (1988) reportedheavy predation on hatchery smolts and wild Atlan -tic salmon and brown trout from the river Bush inNorthern Ireland by the great cormorant Phalacroco-rax carbo. This suggested a link between the re leaseof captive bred smolts (a proxy for farm escapes), theattraction of increased numbers of these predatorybirds to the river, and increased predation on the

437Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

river’s wild Atlantic salmon and brown trout. Simi-larly, Hamoutene et al. (2018) conducted experimen-tal releases and tracking of aquaculture Atlanticsalmon near cage sites in southern Newfoundland,Canada. They found that most released fish were notdetected beyond a few weeks of re lease, with tem-perature and movement data supporting predationas a cause. Increased predation of wild salmon smoltsor adults near sea cages could therefore drive demo-graphic decline or potentially act as a se lective agentif predators cued on size, behaviour, or other traits.Moreover, rates of predation may be higher for indi-viduals already experiencing infections, such as sealice (see Section 2.2). Krkošek et al. (2011b) reportedexperimental evidence that predators selectively con-suming infected prey which could simultaneouslyimpose predation associated impacts and amplify disease or parasite associated selection and mortality.

2.4. Ecological and non-reproductive geneticeffects through competitive interactions

Ecological and non-reproductive genetic effectshave also been suggested via evidence for competi-tive interactions among farm and wild salmon. Thesecompetitive effects could be the result of ecologicalinteractions among wild, farm escaped and hybridoffspring involving differences in behaviour amongcross types such as in aggression, dominance, riskproneness, feeding/foraging activity. And as such,competition associated with these behavioral differ-ences may influence survival and the selective envi-ronments experienced by wild fish. Given the clearoverlap in habitat use-, and evidence for density de-pendence, these seem most likely to take place infreshwater during the juvenile stage (Table 1D). Thishas been illustrated by the work of Fleming et al.(2000), who released sexually mature farm and wildAtlantic salmon into the River Imsa in Norway. De -spite the farm fish achieving less than one-third of thebreeding success compared to wild fish, there wasevidence of resource competition and competitivedisplacement, as the productivity of the wild fish wasdepressed by more than 30%. Fleming et al. (2000)concluded that invasions of farm fish have the poten-tial for impacting wild population productivity bothvia changes to locally adaptive traits as well as re -ductions in genetic diversity. Skaala et al. (2012) doc-umented similar effects in another natural system inNorway. These authors compared the performance offarm, wild, and hybrid Atlantic salmon and suggestedthat overlap in diets and competitions can impact

wild productivity, which could reduce genetic varia-tion in wild populations. Supporting this hypo thesis,Robertsen et al. (2019) demonstrated that the pres-ence of farmed−wild hybrids reduced the survival ofwild half-sibs under semi-natural conditions. There isalso clear evidence that escaped farmed salmon cancompete for spawning habitats and may superimposeredds on top of those of wild Atlantic salmon (Webbet al. 1991, 1993a,b, Fleming et al. 1996). Such super-imposition of redds could af fect both spawning timeand location of wild fish, as well as the growth andsurvival of wild offspring. Overall, it seems highlyprobable that increased competition can result inchanges to the selective landscape experienced bywild individuals and in reductions in population size.

3. QUANTIFYING GENETIC EFFECTS OF NON-REPRODUCTIVE ECOLOGICAL INTERACTIONS

The studies reviewed above demonstrate strongpotential for non-reproductive genetic interactions tooc cur in wild populations. However, quantifyingthese interactions between wild populations anddomestic strains remains a major challenge, particu-larly when hybridization is occurring (i.e. directgenetic interactions). Dramatic increases in DNA se -quencing capacity over the last decade present newopportunities for the use of genomic tools to quantifythe impacts of net pen aquaculture on wild popula-tions. Non-reproductive genetic interactions repre-sent a special, more complex challenge, and the util-ity of genetic and genomic tools to resolve thesegenetic interactions will depend on the route andgenomic scale of impact. That said, a large body ofliterature has been produced in recent years on theuse of genetic/genomic tools to quantify both adap-tive diversity and neutral diversity and effective pop-ulation size or changes therein. As such, a clearopportunity exists to apply genetic and genomicmethods to quantify these impacts.

3.1. Detecting changes in adaptive diversity

In the context of impacts due to changes in the se -lective landscape driven by ecological change, geno -mic change could be associated with a single gene, ormany genes (i.e. polygenic). Genetic and genomictools are increasingly being used to quantify the magnitude of natural selection in the wild (Vitti et al.2013) and many approaches have been developed(Table 2A). One of the best approaches to quantify

438 Aquacult Environ Interact 12: 429–445, 2020

439Bradbury et al.: Genetic impacts of non-reproductive ecological interactionsT

able

2. S

um

mar

y of

ava

ilab

le g

enet

ic a

nd

gen

omic

met

hod

s to

eva

luat

e n

on-r

epro

du

ctiv

e g

enet

ic in

tera

ctio

ns

Met

hod

Com

par

ison

Sta

tist

ics/

test

sR

efer

ence

(A)

Ch

ang

es i

n a

dap

tive

div

ersi

tyT

ime-

seri

es a

nal

ysis

Ch

ang

es in

alle

le f

req

uen

cyE

mp

iric

al li

kel

ihoo

d r

atio

tes

t (E

LR

)F

eder

et

al. (

2014

)C

han

ges

in a

llele

fre

qu

ency

Fre

qu

ency

incr

emen

t te

st (

FIT

)F

eder

et

al. (

2014

)

Tem

por

al c

omp

aris

ons,

pre

- vs

. pos

t-im

pac

tC

han

ges

in a

llele

fre

qu

enci

esP

rin

cip

al c

omp

onen

t an

alys

is, o

utl

ier

det

ecti

on, g

enet

ic d

iffe

ren

tiat

ion

(F

ST)

Bit

ter

et a

l. (2

019)

Tem

por

al c

omp

aris

ons,

pre

- vs

. pos

t-im

pac

tC

han

ges

in a

llele

fre

qu

enci

es in

resp

onse

to

size

-sel

ecti

on g

rad

ien

ts%

pol

ymor

ph

ism

, nu

cleo

tid

e d

iver

sity

, &al

lele

fre

qu

ency

sh

ifts

(co

ntr

ols

vs.

exp

erim

enta

l sam

ple

s)

Th

erk

ildse

n e

t al

. (20

19)

Dom

esti

c an

cest

ry e

stim

atio

n u

nd

er d

iffe

ren

tst

ock

ing

inte

nsi

ties

Rel

atio

nsh

ip b

etw

een

dom

esti

c an

cest

ryan

d r

ecom

bin

atio

n r

ate

at d

iffe

ren

tg

enom

ic s

cale

s

Lei

twei

n e

t al

. (20

19)

Ou

tlie

r d

etec

tion

Loc

us-

spec

ific

com

par

ison

of

pos

teri

orp

rob

abili

ties

of

mod

els

wit

h a

nd

wit

hou

tse

lect

ion

FS

Tco

effi

cien

t &

an

d B

ayes

fac

tor

scor

esF

oll &

Gag

gio

tti (

2008

)

Ou

tlie

r d

etec

tion

Tes

ts o

f n

eutr

alit

y b

ased

on

pri

nci

pal

com

pon

ents

an

alys

isM

ahal

anob

is d

ista

nce

Lu

u e

t al

. (20

17)

Imp

acte

d v

s. n

on-i

mp

acte

dS

ign

atu

res

of s

elec

tion

th

at c

ovar

y w

ith

envi

ron

men

tal s

tres

sor

(e.g

. pol

luti

on)

FS

T, p

opu

lati

on b

ran

ch s

tati

stic

, dif

fer-

ence

s in

nu

cleo

tid

e d

iver

sity

alo

ng

20-

kilo

bas

e sl

idin

g w

ind

ow

Ozi

olor

et

al. (

2019

)

Imp

acte

d v

s. n

on-i

mp

acte

dS

ign

atu

res

of s

elec

tion

ass

ocia

ted

wit

hen

viro

nm

enta

l str

esso

rF

ST

outl

ier

(FD

IST

2)D

ayan

et

al. (

2019

)

Gen

ome-

wid

e as

soci

atio

n s

tud

ies

Pol

ygen

ic a

ssoc

iati

ons

wit

h p

opu

lati

ond

eclin

e in

volv

ing

gen

omic

reg

ion

sre

late

d t

o m

etab

olis

m, d

evel

opm

enta

l &p

hys

iolo

gic

al p

roce

sses

Ch

ang

e in

μ(s

ign

atu

re o

f se

lect

ive

swee

ps)

bet

wee

n d

eclin

ing

an

d n

on-

dec

linin

g p

opu

lati

on s

tatu

s of

Atl

anti

csa

lmon

; Red

un

dan

cy a

nal

ysis

(R

DA

) fo

rd

etec

tion

of

outl

iers

, pol

ygen

ic r

isk

sco

res

Leh

ner

t et

al.

(201

9)

Sof

t se

lect

ive

swee

ps

Iden

tifi

cati

on o

f n

ew a

llele

s to

inte

rmed

i-at

e fr

equ

ency

ag

ain

st a

bac

kg

rou

nd

of

un

usu

ally

lon

g h

aplo

typ

es o

f lo

wn

ucl

eoti

de

div

ersi

ty

Inte

gra

ted

hap

loty

pe

scor

es (

iHS

)V

oig

ht

et a

l. (2

006)

Sof

t se

lect

ive

swee

ps

Iden

tifi

cati

on o

f se

lect

ed a

llele

s n

eari

ng

or h

avin

g a

chie

ved

fix

atio

n in

on

ep

opu

lati

on b

ut

that

rem

ain

s p

olym

orp

hic

in t

he

wid

er g

rou

p o

f p

opu

lati

ons.

Ext

end

ed c

ross

pop

ula

tion

hap

loty

pe

hom

ozyg

osit

y (X

P-E

HH

)S

abet

i et

al. (

2007

)

Sof

t se

lect

ive

swee

ps

Det

ecti

on o

f p

osit

ive

sele

ctio

n a

ctin

g t

oin

crea

se h

aplo

typ

e h

omoz

ygos

ity;

com

bin

es d

istr

ibu

tion

of

frag

men

tle

ng

ths

bet

wee

n m

uta

tion

s an

d n

um

ber

of s

egre

gat

ing

sit

es b

etw

een

all

pai

rs o

fch

rom

osom

es; r

atio

of

hap

loty

pe

hom

ozyg

osit

y fo

r d

eriv

ed &

an

cest

ral

alle

les.

Nu

mb

er o

f se

gre

gat

ing

sit

es b

y le

ng

th(n

SL

); s

imila

r to

iHS

bu

t (1

) a

gen

etic

map

is n

ot r

equ

ired

an

d (

2) m

ore

rob

ust

to

reco

mb

inat

ion

an

d/o

r m

uta

tion

rat

eva

riat

ion

Fer

rer-

Ad

met

lla e

t al

.(2

014)

Tab

le c

onti

nu

ed o

n n

ext

pag

e

the presence of selection is either the comparisonof representative pre- and post-impact gene ticsamples in the absence of hybridization or theexamination of situations with the capacity toquantify and correct for signatures of recent orcurrent hybridization (Leit wein et al. 2019). Fortime series analysis of changes in allele fre-quency associated with selection, differentiationmeasures such as the fixation index (FST) arecom mon ly used, and several tests have been re -cently proposed using bi-allelic loci, includingthe empirical likelihood ratio test (ELRT) and thefrequency increment test (FIT) (Feder et al. 2014).Recent temporal comparisons of natural selectionin ecological, climate adaptation, and fishery-impact studies have re vealed detectable increasesin genomic differentiation over even short time-frames (e.g. 1 to 4 generations; Bitter et al. 2019,Leitwein et al. 2019, Therkildsen et al. 2019), in-dicating genomic tools show high power to de-tect changes in natural se lection when recentpre-impact baselines are available. Where repli-cate temporal comparisons across sites can bemade, this may allow uncovering parallel pat-terns and non-parallel signatures of adaptation.Knowledge of pre-impact genomic variationacross replicates could quantify both the sourceand magnitude of non-reproductive genetic im-pacts; sites with similar starting genomic varia-tion are more likely to show parallel responses,unless source or strength of selection differs.

In the absence of pre-impact samples, tradi-tional tests for the presence of outliers (e.g. Foll& Gaggiotti 2008, Luu et al. 2017), trait asso -ciations, or selective sweeps (e.g. Nielsen 2005)may be applied using genome-wide polymor-phism data, though the ability to attribute agiven impact to these loci may be problematic.Similar to pre- and post-impact temporal com-parisons, tests for genomic differentiation usingmetrics such as FST between sites with differinglevels of exposure to stressors can be used todetect the magnitude and location of genomicchange between these impacted and pristinesites (e.g. Dayan et al. 2019, Oziolor et al.2019). Genome-wide association and genome–environment association methods also showpromise in measuring aquaculture impacts, buthave traditionally been used to estimate corre-lations be tween genomic variants and trait orenvironmental variation (Rellstab et al. 2015,Santure & Garant 2018). A recent genomicstudy by Lehnert et al. (2019) instead used

440 Aquacult Environ Interact 12: 429–445, 2020

Tab

le 2

(con

tin

ued

)

Met

hod

Com

par

ison

Sta

tist

ics/

Tes

tsR

efer

ence

Mac

hin

e le

arn

ing

Cor

rela

tes

of h

abit

at/e

nvi

ron

men

tal

vari

able

s w

ith

ob

serv

ed g

enet

ic s

tru

c-tu

re

Ran

dom

For

est;

PC

A lo

adin

gs;

ou

tlie

rd

etec

tion

Syl

vest

er e

t al

. (20

18a)

Mac

hin

e le

arn

ing

Det

ecti

on o

f lo

ci o

f sm

all p

hen

otyp

icef

fect

on

a k

ey li

fe-h

isto

ry v

aria

ble

(e.

g.

run

tim

ing

) ac

ross

mu

ltip

le p

opu

lati

ons

Ran

dom

for

est;

ou

tlie

r d

etec

tion

; PC

AB

rieu

c et

al.

(201

5)

(B)

Ch

ang

es i

n n

eutr

al d

iver

sity

or

effe

ctiv

e p

op

ula

tio

n s

ize

Eff

ecti

ve p

opu

lati

on s

ize

Sin

gle

-sam

ple

met

hod

bas

ed o

n li

nk

age

dis

equ

ilib

riu

m t

o es

tim

ate

effe

ctiv

ep

opu

lati

ons

size

Con

tem

por

ary

Ne

Wap

les

& D

o (2

010)

,W

aple

s et

al.

(201

6)

Eff

ecti

ve p

opu

lati

on s

ize

Sin

gle

-sam

ple

met

hod

to

esti

mat

ech

ang

es in

con

tem

por

ary

Ne

by

com

par

-in

g li

nk

age

dis

equ

ilib

riu

m e

stim

ates

wit

h r

ecom

bin

atio

n r

ates

est

imat

ed f

rom

ph

ysic

al li

nk

age

or g

enom

ic p

osit

ion

Con

tem

por

ary

Ne

esti

mat

es a

t va

riou

sti

mes

in t

he

pas

tH

olle

nb

eck

et

al. (

2016

)

Eff

ecti

ve p

opu

lati

on s

ize

Ap

plic

atio

n o

f H

olle

nb

eck

et

al. (

2016

)fo

r ra

ng

e-w

ide

pop

ula

tion

s of

Atl

anti

csa

lmon

an

d a

ssoc

iati

ons

of g

enom

icre

gio

ns

to d

eclin

e st

atu

s

Con

tem

por

ary

Ne

esti

mat

es o

ver

tim

eL

ehn

ert

et a

l. (2

019)

decline status as the trait in genome-wide associationand uncovered polygenic associations with popula-tion decline and variation in immune and develop-mental genes. This ap proach could be further refinedin future studies by incorporating continuous meas-ures of aquaculture exposure such as magnitude ofescape, site proximity, or pathogen load.

Rapid evolutionary change is often associated withselection on standing genetic variation (‘soft sweeps’)rather than new mutations (Messer et al. 2016, Her-misson & Pennings 2017). Methods that utilize differ-ences in frequency and diversity of haplotypes suchas integrated haplotype score (iHS; Voight et al.2006), extended cross population haplotype homozy-gosity (XP-EHH; Sabeti et al. 2007), and number ofsegregating sites by length (nSL; Ferrer-Admetlla etal. 2014) can identify signatures of soft selectivesweeps. Identification of sweep signatures that areex clusive to aquaculture-impacted populations mayprovide an additional way of both validating geno -mic changes induced by non-reproductive geneticim pacts and uncovering implicated target genes.Machine learning approaches have also shown prom-ise in identifying subtle signatures of environment(Sylvester et al. 2018a), trait associations (Brieuc et al.2015), and selective sweep signatures (Kern &Schrider 2018). These provide additional re searchareas for future studies into the gene tic impacts ofaquaculture exposure that may not be de tected bytraditional statistical approaches. Lastly, gene ontol-ogy (Rivals et al. 2007) and gene set (Daub et al.2017) enrichment methods can be used to character-ize functional impacts and parallel responses at bio-logical levels above changes at individual genes(Jacobs et al. 2020) and can help clarify potential tar-gets of selection from aquaculture interactions.

3.2. Detecting changes in neutral diversity oreffective population size

Genomic approaches can also be applied in thecontext of resolving a loss of diversity due to demo-graphic declines associated with non-reproductivegenetic impacts and applied to quantify genome-wide trends in diversity over time or estimate trendsin the effective population size (Table 2B; see Waples& Do 2010). Large genomic datasets offer new oppor-tunities for enhanced estimates of effective popula-tion size (Waples et al. 2016) as well as retrospectiveestimates of changes in effective population size overtime (e.g. Hollenbeck et al. 2016). For example, B.Watson (pers. comm.) evaluated the performance of

estimates of effective population size (Ne) using largegenomic datasets to assess and approximate popula-tion declines. This was used to establish a genomicbaseline to detect non-reproductive ge netic interac-tions in southern Newfoundland Atlantic salmonpopulations following the use of largely sterile Atlan -tic salmon in aquaculture. Their results suggest thatlarge genomic datasets (≥1000 SNPs) were able todetect population declines significantly earlier, andwith increased accuracy, than small genetic or geno -mic datasets (25 microsatellites or 100 SNPs). How-ever, monitoring using effective size requires sam-ples from multiple time points, which is not alwayspossible. As an alternative, Hollenbeck et al. (2016)present a method that uses linkage information tobin loci by rates of recombination and reconstructtrends in Ne decades into the past. Lehnert et al.(2019) applied this method to Atlantic salmon acrossthe North Atlantic and estimated that 60% of all pop-ulations have declined in recent decades. Finally,molecular approaches to mark-recapture abundanceestimation (i.e. CKMR, Bravington et al. 2016) alsooffer the potential to quantify changes in populationsize over time and have been used in marine andfreshwater fish species (Bravington et al. 2016,Waples et al. 2018, Ruzzante et al. 2019). Such ap -proaches could be used to quantify population trendsin effective size in the absence of assessment dataand monitor for ecological and non-reproductivegenetic interactions in future.

4. CONCLUSIONS

Ultimately, despite an abundance of relevant andinformative re search, the relative importance of hybrid -ization and non-reproductive genetic inter actions be-tween domestic individuals and wild populations re-mains largely unresolved. Nonetheless, the literaturesuggests that ecological interactions arising fromsalmon aquaculture have the realistic potential to re-sult in substantial genetic change in wild salmon pop-ulations, as well as other species. It is worth notingthat, at present, there is a significant knowledge gapre garding the non-reproductive genetic impacts of in -creased predation or competition due to salmon aqua-culture on wild populations. Fortunately, recent ad-vances in genetic and genomic methods present anew scope for quantifying these impacts. However,careful experimental design and pre-impact compar-isons will in most cases be needed to accurately attrib-ute any genetic change to non-reproductive geneticinteractions with salmon aquaculture activities.

441Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

Future research should explore the sensitivitiesand power of these approaches to detect changes ingenetic diversity and character over time. Given thatboth reproductive and non-reproductive interactionsco-occur within the native range of Atlantic salmon,there may be benefit to focus studies on instanceswhere interbreeding is unlikely or impossible. Thiscould involve the study of ecological and geneticimpacts in other species such as Pacific salmon speciesor in Atlantic salmon in regions where sterility isemployed as a containment or mitigation measure.Alternatively, genomic approaches could poten-tially be used to disentangle reproductive and non- re productive interactions from indirect interactionsbased on the identification of hybrids, introgressedancestry blocks, or signatures of selection.

Our review suggests that non-reproductive geneticinteractions represent both a broad reaching andlargely unresolved source of genetic impact on wildpopulations exposed to Atlantic salmon aquacultureactivities. Thus, further study is urgently needed tosupport an integrated understanding of aquaculture−ecosystem interactions, their implications for ecosys-tem stability, and the identification of potential path-ways of effect. This information will be essential tothe development of potential mitigation and manage-ment strategies.

Acknowledgements. This review was conducted as part ofthe ICES Working Group on Genetic Applications to Fish-eries and Aquaculture between 2018 and 2020. Fundingwas provided through the Fisheries and Oceans Program forAquaculture Regulatory Research. This work also benefitedgreatly from a 3-year Canada-EU Galway Statement for theTransatlantic Ocean Research Alliance Working Group onmodelling genetic interactions among wild and farmescaped Atlantic salmon in the North Atlantic. The authorsthank Brian Dempson and several anonymous reviewers forcomments on this manuscript.

LITERATURE CITED

Bitter MC, Kapsenberg L, Gattuso JP, Pfister CA (2019)Standing genetic variation fuels rapid adaptation toocean acidification. Nat Commun 10: 5821

Bjørn PA, Finstad B, Kristoffersen R (2001) Salmon lice infec-tion of wild sea trout and Arctic char in marine and fresh-waters: the effects of salmon farms. Aquacult Res 32: 947−962

Bøhn T, Gjelland KØ, Serra-Llinares RM, Finstad B and oth-ers (2020) Timing is everything: survival of Atlanticsalmon Salmo salar postsmolts during events of highsalmon lice densities. J Appl Ecol 57: 1149−1160

Bolstad GH, Hindar K, Robertsen G, Jonsson B and others(2017) Gene flow from domesticated escapes alters thelife history of wild Atlantic salmon. Nat Ecol Evol 1: 0124

Børretzen Fjørtoft H Nilsen F, Besnier F, Espedal PG and

others (2020) Aquaculture-driven evolution: distributionof pyrethroid resistance in the salmon louse throughoutthe North Atlantic in the years 2000−2017. ICES J MarSci 77: 1806−1815

Bourret V, O’Reilly PT, Carr JW, Berg PR, Bernatchez L(2011) Temporal change in genetic integrity suggestsloss of local adaptation in a wild Atlantic salmon (Salmosalar) population following introgression by farmedescapees. Heredity 106: 500−510

Bradbury IR, Duffy S, Lehnert SJ, Jóhannsson R and others(2020) Model-based evaluation of the genetic impacts offarm-escaped Atlantic salmon on wild populations.Aquacult Environ Interact 12: 45−59

Bravington MV, Grewe PM, Davies CR (2016) Absoluteabundance of southern bluefin tuna estimated by close-kin mark-recapture. Nat Commun 7: 13162

Brieuc MSO, Ono K, Drinan DP, Naish KA (2015) Integrationof Random Forest with population-based outlier analysesprovides insight on the genomic basis and evolution ofrun timing in Chinook salmon (Oncorhynchus tsha wyt -scha). Mol Ecol 24: 2729−2746

Bürger R, Lynch M (1995) Evolution and extinction in achanging environment: a quantitative genetic analysis.Evolution 49: 151−163

Callier MD, Byron CJ, Bengtson DA, Cranford PJ and others(2018) Attraction and repulsion of mobile wild organismsto finfish and shellfish aquaculture: a review. Rev Aqua -cult 10: 924−949

Connors BM, Krkošek M, Ford J, Dill LM (2010) Cohosalmon productivity in relation to salmon lice from in -fected prey and salmon farms. J Appl Ecol 47: 1372−1377

Consuegra S, de Eyto E, McGinnity P, Stet RJM, Jordan WC(2011) Contrasting responses to selection in class I andclass IIα major histocompatibility-linked markers insalmon. Heredity 107: 143

Correa K, Lhorente JP, Bassini L, López ME and others (2017)Genome wide association study for resistance to Caligusrogercresseyi in Atlantic salmon (Salmo salar L.) using a50K SNP genotyping array. Aquaculture 472: 61−65

Coughlan J, McGinnity P, O’Farrell B, Dillane E and others(2006) Temporal variation in an immune response gene(MHC I) in anadromous Salmo trutta in an Irish riverbefore and during aquaculture activities. ICES J Mar Sci63: 1248−1255

Cross TF, King J (1983) Genetic effects of hatchery rearingin Atlantic salmon. Aquaculture 33: 33−40

Daub JT, Moretti S, Davydov II, Excoffier L, Robinson-Rechavi M (2017) Detection of pathways affected by pos-itive selection in primate lineages ancestral to humans.Mol Biol Evol 34: 1391−1402

Dayan DI, Du X, Baris TZ, Wagner DN, Crawford DL, Olek-siak MF (2019) Population genomics of rapid evolution innatural populations: polygenic selection in response topower station thermal effluents. BMC Evol Biol 19: 61

de Eyto E, McGinnity P, Consuegra S, Coughlan J and oth-ers (2007) Natural selection acts on Atlantic salmonmajor histocompatibility (MH) variability in the wild.Proc R Soc B 274: 861−869

de Eyto E, McGinnity P, Huisman J, Coughlan J and others(2011) Varying disease-mediated selection at differentlife-history stages of Atlantic salmon in fresh water. EvolAppl 4: 749−762

Denholm SJ, Hoyle AS, Shinn AP, Paladini G, Taylor NGH,Norman RA (2016) Predicting the potential for naturalrecovery of Atlantic salmon (Salmo salar L.) populations

442 Aquacult Environ Interact 12: 429–445, 2020

following the introduction of Gyrodactylus salaris Malm-berg, 1957 (Monogenea). PLOS ONE 11: e0169168

DFO (2016) Proposed use of European-strain triploid Atlanticsalmon in Marine cage aquaculture in Placentia Bay, NL.DFO Can Sci Advis Sec Sci Resp 2016/034. https:// waves-vagues.dfo-mpo.gc.ca/Library/40621248.pdf

DFO (2018) Review of the environmental impact statementfor the Placentia Bay Atlantic salmon aquaculture project.DFO Can Sci Advis Sec Sci Resp 2018. https://waves-vagues.dfo-mpo.gc.ca/Library/40738899.pdf

Dionne M, Miller KM, Dodson JJ, Caron F, Bernatchez L(2007) Clinal variation in MHC diversity with tempera-ture: evidence for the role of host-pathogen interactionon local adaptation in Atlantic salmon. Evolution 61: 2154−2164

Dionne M, Miller KM, Dodson JJ, Bernatchez L (2009) MHCstanding genetic variation and pathogen resistance inwild Atlantic salmon. Philos Trans R Soc B 364: 1555−1565

Diserud OH, Fiske P, Sægrov H, Urdal K and others (2019)Escaped farmed Atlantic salmon in Norwegian riversduring 1989−2013. ICES J Mar Sci 76: 1140−1150

Feder AF, Kryazhimskiy S, Plotkin JB (2014) Identifying sig-natures of selection in genetic time series. Genetics 196: 509−522

Ferrer-Admetlla A, Liang M, Korneliussen T, Nielsen R (2014)On detecting incomplete soft or hard selective sweepsusing haplotype structure. Mol Biol Evol 31: 1275−1291

Fjelldal PG, Hansen TJ, Karlsen Ø, Wright DW (2019)Effects of laboratory salmon louse infection on Arcticchar osmoregulation, growth and survival. ConservPhysiol 7: coz072

Fleming IA, Jonsson B, Gross MR (1994) Phenotypic diver-gence of sea-ranched, farmed, and wild salmon. Can JFish Aquat Sci 51: 2808−2824

Fleming IA, Jonsson B, Gross MR, Lamberg A (1996) Anexperimental study of the reproductive behaviour andsuccess of farmed and wild Atlantic salmon (Salmo salar).J Appl Ecol 33: 893−905

Fleming IA, Hindar K, Mjølnerød IB, Jonsson B, Balstad T,Lamberg A (2000) Lifetime success and interactions offarm salmon invading a native population. Proc R Soc B267: 1517−1524

Foll M, Gaggiotti O (2008) A Genome-scan method to iden-tify selected loci appropriate for both dominant andcodominant markers: a Bayesian perspective. Genetics180: 977−993

Ford JS, Myers RA (2008) A global assessment of salmonaquaculture impacts on wild salmonids. PLOS Biol 6: e33

Forseth T, Barlaup BT, Finstad B, Fiske P and others (2017)The major threats to Atlantic salmon in Norway. ICES JMar Sci 74: 1496−1513

Garcia de Leaniz C, Fleming IA, Einum S, Verspoor E andothers (2007) A critical review of adaptive genetic varia-tion in Atlantic salmon: implications for conservation.Biol Rev Camb Philos Soc 82: 173−211

Garseth ÅH, Ekrem T, Biering E (2013) Phylogenetic evi-dence of long distance dispersal and transmission ofpiscine reovirus (PRV) between farmed and wild Atlanticsalmon. PLOS ONE 8: e82202

Gilbey J, Verspoor E, Mo TA, Sterud E and others (2006)Identification of genetic markers associated with Gyro-dactylus salaris resistance in Atlantic salmon (Salmosalar). Dis Aquat Org 71: 119−129

Glover KA, Pertoldi C, Besnier F, Wennevik V, Kent M,Skaala Ø (2013) Atlantic salmon populations invaded byfarmed escapees: quantifying genetic introgression witha Bayesian approach and SNPs. BMC Genet 14: 74

Glover KA, Solberg MF, McGinnity P, Hindar K and others(2017) Half a century of genetic interaction betweenfarmed and wild Atlantic salmon: status of knowledgeand unanswered questions. Fish Fish 18: 890−927

Gomulkiewicz R, Holt RD (1995) When does evolution by nat-ural selection prevent extinction? Evolution 49: 201−207

Hamoutene D, Cote D, Marshall K, Donnet S and others(2018) Spatial and temporal distribution of farmed Atlanticsalmon after experimental release from sea cage sites inNewfoundland (Canada). Aquaculture 492: 147−156

Handeland SO, Björnsson BT, Arnesen AM, Stefansson SO(2003) Seawater adaptation and growth of post-smoltAtlantic salmon (Salmo salar) of wild and farmed strains.Aquaculture 220: 367−384

Hermisson J, Pennings PS (2017) Soft sweeps and beyond: understanding the patterns and probabilities of selectionfootprints under rapid adaptation. Methods Ecol Evol 8: 700−716

Hollenbeck CM, Portnoy DS, Gold JR (2016) A method fordetecting recent changes in contemporary effective pop-ulation size from linkage disequilibrium at linked andunlinked loci. Heredity 117: 207

Jacobs A, Carruthers M, Yurchenko A, Gordeeva NV andothers (2020) Parallelism in eco-morphology and geneexpression despite variable evolutionary and genomicbackgrounds in a Holarctic fish. PLOS Genet 16: e1008658

Johnsen BO, Jensen AJ (1991) The Gyrodactylus story inNorway. Aquaculture 98: 289−302

Johnsen BO, Jensen AJ (1994) The spread of furunculosis insalmonids in Norwegian rivers. J Fish Biol 45: 47−55

Karbowski CM, Finstad B, Karbowski N, Hedger RD (2019)Sea lice in Iceland: assessing the status and current impli-cations for aquaculture and wild salmonids. AquacultEnviron Interact 11: 149−160

Karlsson S, Moen T, Lien S, Glover KA, Hindar K (2011)Generic genetic differences between farmed and wildAtlantic salmon identified from a 7K SNP-chip. Mol EcolResour 11(Suppl 1): 247−253

Karlsson S, Diserud OH, Fiske P, Hindar K (2016) Wide-spread genetic introgression of escaped farmed Atlanticsalmon in wild salmon populations. ICES J Mar Sci 73: 2488−2498

Karlsson S, Bolstad GH, Hansen H, Jansen PA, Moen T,Noble LR (2020) The potential for evolution of resistanceto Gyrodactylus salaris in Norwegian Atlantic salmon.NINA Report 1812, Norwegian Institute for NatureResearch, Trondheim

Kennedy GJA, Greer JE (1988) Predation by cormorantsPhala cro corax carbo (L.), on an Irish river. Aquacult FishManage 19: 159−170

Kern AD, Schrider DR (2018) diploS/HIC: an updated ap -proach to classifying selective sweeps. G3 (Bethesda) 8: 1959−1970

Keyser F, Wringe BF, Jeffery NW, Dempson JB, Duffy S,Bradbury IR (2018) Predicting the impacts of escapedfarmed Atlantic salmon on wild salmon populations. CanJ Fish Aquat Sci 75: 506−512

Kirkpatrick M, Barton NH (1997) Evolution of species’range. Am Nat 150: 1−23

Kjærner-Semb E, Ayllon F, Furmanek T, Wennevik V andothers (2016) Atlantic salmon populations reveal adap-

443Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

tive divergence of immune related genes — a duplicatedgenome under selection. BMC Genomics 17: 610

Krkošek M, Lewis MA, Volpe JP (2005) Transmission dyna -mics of parasitic sea lice from farm to wild salmon. ProcBiol Sci 272: 689−696

Krkošek M, Ford JS, Morton A, Lele S, Myers RA, Lewis MA(2007) Declining wild salmon populations in relation toparasites from farm salmon. Science 318: 1772−1775

Krkošek M, Connors BM, Morton A, Lewis MA, Dill LM,Hilborn R (2011a) Effects of parasites from salmon farmson productivity of wild salmon. Proc Natl Acad Sci USA108: 14700−14704

Krkošek M, Connors BM, Ford H, Peacock S and others(2011b) Fish farms, parasites, and predators: implicationsfor salmon population dynamics. Ecol Appl 21: 897−914

Lehnert SJ, Kess T, Bentzen P, Kent MP and others (2019)Genomic signatures and correlates of widespread popu-lation declines in salmon. Nat Commun 10: 2996

Leitwein M, Cayuela H, Ferchaud AL, Normandeau É, Gag-naire PA, Bernatchez L (2019) The role of recombinationon genome-wide patterns of local ancestry exemplifiedby supplemented brook charr populations. Mol Ecol 28: 4755−4769

Luu K, Bazin E, Blum MGB (2017) pcadapt: an R package toperform genome scans for selection based on principalcomponent analysis. Mol Ecol Resour 17:67–77

Madhun AS, Karlsbakk E, Isachsen CH, Omdal LM andothers (2015) Potential disease interaction reinforced: double-virus-infected escaped farmed Atlantic salmon,Salmo salar L., recaptured in a nearby river. J Fish Dis38: 209−219

Madhun AS, Isachsen CH, Omdal LM, Einen ACB and others(2018) Prevalence of piscine orthoreovirus and salmonidalphavirus in sea-caught returning adult Atlantic salmon(Salmo salar L.) in northern Norway. J Fish Dis 41: 797−803

McGinnity P, Prodöhl P, Ferguson A, Hynes R and others(2003) Fitness reduction and potential extinction of wildpopulations of Atlantic salmon, Salmo salar, as a result ofinteractions with escaped farm salmon. Proc R Soc B 270: 2443−2450

Messer PW, Ellner SP, Hairston NG Jr (2016) Can populationgenetics adapt to rapid evolution? Trends Genet 32: 408−418

Miller KM, Li S, Kaukinen KH, Ginther N and others (2011)Genomic signatures predict migration and spawningfailure in wild Canadian salmon. Science 331: 214−217

Morris MRJ, Fraser DJ, Heggelin AJ, Whoriskey FG, CarrJW, O’Neil SF, Hutchings JA (2008) Prevalence andrecurrence of escaped farmed Atlantic salmon (Salmosalar) in eastern North American rivers. Can J FishAquat Sci 65: 2807−2826

Morton A, Routledge R, Hrushowy S, Kibenge M, Kibenge F(2017) The effect of exposure to farmed salmon on piscineorthoreovirus infection and fitness in wild Pacific salmonin British Columbia, Canada. PLOS ONE 12: e0188793

Nekouei O, Vanderstichel R, Thakur K, Arriagada G andothers (2018) Association between sea lice (Lepeoph-theirus salmonis) infestation on Atlantic salmon farmsand wild Pacific salmon in Muchalat Inlet, Canada. SciRep 8: 4023

Nielsen R (2005) Molecular signatures of natural selection.Annu Rev Genet 39: 197−218

Nylund A, Brattespe J, Plarre H, Kambestad M, Karlsen M(2019) Wild and farmed salmon (Salmo salar) as reser-voirs for infectious salmon anaemia virus, and the impor-

tance of horizontal- and vertical transmission. PLOSONE 14: e0215478

Oziolor EM, Reid NM, Yair S, Lee KM and others (2019)Adaptive introgression enables evolutionary rescue fromextreme environmental pollution. Science 364: 455−457

Peacock SJ, Krkošek M, Proboszcz S, Orr C, Lewis MA(2013) Cessation of a salmon decline with control of par-asites. Ecol Appl 23: 606−620

Price MH, Proboszcz SL, Routledge RD, Gottesfeld AS, OrrC, Reynolds JD (2011) Sea louse infection of juvenilesockeye salmon in relation to marine salmon farms onCanada’s west coast. PLOS ONE 6: e16851

Pritchard VL, Mäkinen H, Vähä JP, Erkinaro J, Orell P,Primmer CR (2018) Genomic signatures of fine-scalelocal selection in Atlantic salmon suggest involvement ofsexual maturation, energy homeostasis and immunedefence-related genes. Mol Ecol 27: 2560−2575

Rellstab C, Gugerli F, Eckert AJ, Hancock AM, HoldereggerR (2015) A practical guide to environmental associationanalysis in landscape genomics. Mol Ecol 24: 4348−4370

Rivals I, Personnaz L, Taing L, Potier MC (2007) Enrichmentor depletion of a GO category within a class of genes: which test? Bioinformatics 23: 401−407

Robertsen G, Reid D, Einum S, Aronsen T and others (2019)Can variation in standard metabolic rate explain context-dependent performance of farmed Atlantic salmon off-spring? Ecol Evol 9: 212−222

Robledo D, Gutiérrez AP, Barría A, Lhorente JP, HoustonRD, Yáñez JM (2019) Discovery and functional annota-tion of quantitative trait loci affecting resistance to sealice in Atlantic salmon. Front Genet 10: 56

Ruzzante DE, McCracken GR, Førland B, MacMillan J andothers (2019) Validation of close-kin mark−recapture(CKMR) methods for estimating population abundance.Methods Ecol Evol 10: 1445−1453

Sabeti PC, Varilly P, Fry B, Lohmueller J and others (2007)Genome-wide detection and characterization of positiveselection in human populations. Nature 449: 913−918

Santure AW, Garant D (2018) Wild GWAS — associationmapping in natural populations. Mol Ecol Resour 18: 729−738

Serra-Llinares RM, Bøhn T, Karlsen Ø, Nilsen R and others(2020) Impacts of salmon lice on mortality, marine migra-tion distance and premature return in sea trout. Mar EcolProg Ser 635: 151−168

Shephard S, Gargan P (2017) Quantifying the contributionof sea lice from aquaculture to declining annual returnsin a wild Atlantic salmon population. Aquacult EnvironInteract 9: 181−192

Skaala Ø, Glover KA, Barlaup BT, Svåsand T, Besnier F,Hansen MM, Borgstrøm R (2012) Performance of farmed,hybrid, and wild Atlantic salmon (Salmo salar) families ina natural river environment. Can J Fish Aquat Sci 69: 1994−2006

Skilbrei OT, Heino M, Svåsand T (2015) Using simulatedescape events to assess the annual numbers and destiniesof escaped farmed Atlantic salmon of different life stagesfrom farm sites in Norway. ICES J Mar Sci 72: 670−685

Sylvester EVA, Beiko RG, Bentzen P, Paterson I and others(2018a) Environmental extremes drive population struc-ture at the northern range limit of Atlantic salmon inNorth America. Mol Ecol 27: 4026−4040

Sylvester EVA, Wringe BF, Duffy SJ, Hamilton LC, FlemingIA, Bradbury IR (2018b) Migration effort and wild popu-lation size influence the prevalence of hybridization

444 Aquacult Environ Interact 12: 429–445, 2020

be tween escaped farmed and wild Atlantic salmon.Aquacult Environ Interact 10: 401−411

Therkildsen NO, Wilder AP, Conover DO, Munch SB, Bau-mann H, Palumbi SR (2019) Contrasting genomic shiftsunderlie parallel phenotypic evolution in response tofishing. Science 365: 487−490

Thorstad EB, Finstad B (2018) Impacts of salmon lice ema-nating from salmon farms on wild Atlantic salmon andsea trout. NINA Report 1449, Norwegian Institute forNature Research, Trondheim

Tonteri A, Vasemägi A, Lumme J, Primmer CR (2010) BeyondMHC: signals of elevated selection pressure in Atlanticsalmon (Salmo salar) immune-relevant loci. Mol Ecol 19: 1273−1282

Tsai HY, Hamilton A, Tinch AE, Guy DR and others (2016)Genomic prediction of host resistance to sea lice infarmed Atlantic salmon populations. Genet Sel Evol48:47

Verspoor E, McGinnity P, Bradbury I, Glebe B (2015) Thepotential direct and indirect genetic consequences fornative Newfoundland Atlantic salmon from interbreed-ing with European-origin farm escapes. DFO Can SciAdvis Sec Res Document 2015/030: viii + 36. https:// waves-vagues.dfo-mpo.gc.ca/Library/359696.pdf

Vitti JJ, Grossman SR, Sabeti PC (2013) Detecting naturalselection in genomic data. Annu Rev Genet 47:97–120

Voight BF, Kudaravalli S, Wen X, Pritchard JK (2006) A mapof recent positive selection in the human genome. PLOSBiol 4: e72

Vollset KW, Krontveit RI, Jansen PA, Finstad B and others(2016) Impacts of parasites on marine survival of Atlanticsalmon: a meta-analysis. Fish Fish 17: 714−730

Waples RS, Do C (2010) Linkage disequilibrium estimates ofcontemporary Ne using highly variable genetic markers: a largely untapped resource for applied conservationand evolution. Evol Appl 3: 244−262

Waples RK, Larson WA, Waples RS (2016) Estimating con-temporary effective population size in non-model spe-cies using linkage disequilibrium across thousands ofloci. Heredity 117: 233−240

Waples RS, Grewe PM, Bravington MW, Hillary R, Feutry P(2018) Robust estimates of a high Ne/N ratio in a top mar-ine predator, southern bluefin tuna. Sci Adv 4: eaar7759

Webb JH, Hay DW, Cunningham PD, Youngson AF (1991)The spawning behaviour of escaped farmed and wildadult Atlantic salmon (Salmo salar L.) in a northern Scot-tish river. Aquaculture 98: 97−110

Webb JH, McLaren IS, Donaghy MJ, Youngson AF (1993a)Spawning of farmed Atlantic salmon in the second yearafter their escape. Aquacult Fish Manage 24: 557−561

Webb JH, Youngson AF, Thompson CE, Hay DW, DonaghyMJ, McLaren IS (1993b) Spawning of escaped farmedAtlantic salmon, Salmo salar L., in western and northernScottish rivers: egg deposition by females. Aquacult FishManage 24: 663−670

Wringe BF, Jeffery NW, Stanley RRE, Hamilton LC and oth-ers (2018) Extensive hybridization following a largeescape of domesticated Atlantic salmon in the NorthwestAtlantic. Commun Biol 1: 108

Wringe BF, Anderson EC, Jeffery NW, Stanley RRE, Brad-bury IR (2019) Development and evaluation of SNP pan-els for the detection of hybridization between wild andescaped Atlantic salmon (Salmo salar) in the WestAtlantic. Can J Fish Aquat Sci 76: 695−704

Zueva KJ, Lumme J, Veselov AE, Kent MP, Lien S, PrimmerCR (2014) Footprints of directional selection in wildAtlantic salmon populations: evidence for parasite-dri-ven evolution? PLOS ONE 9: e91672

Zueva KJ, Lumme J, Veselov AE, Kent MP, Primmer CR(2018) Genomic signatures of parasite-driven naturalselection in north European Atlantic salmon (Salmosalar). Mar Genomics 39: 26−38

445Bradbury et al.: Genetic impacts of non-reproductive ecological interactions

Editorial responsibility: Kevin Glover, Bergen, Norway

Submitted: June 8, 2020; Accepted: August 31, 2020Proofs received from author(s): October 12, 2020